Should the analog power supply
be separated from the digital one?

Described as a best practice, often recommended and quite common among audiophiles, is the practice of separating the power lines of audio systems by clearly dividing them:
digital on one side, analog on the other.

But is that really a good idea?

Let’s start with the purpose: What is — or what should be — the purpose of separating power supplies in the audio chain?

If it is true that the components of a music playback system influence one another through the common power line, we need to understand HOW this happens.

We know, in fact, that every power supply draws energy from the electrical grid, and it does so in ways that depend on the type of power supply (we typically distinguish between linear and switching power supplies) and the type of power consumption.

The transformer in every power supply operates bidirectionally: it transfers what happens on the primary side to the secondary side—that is, to the secondary winding. But the opposite is also true: the power draw generated by the load connected to the secondary side is inevitably reflected back to the primary side and from there to the power grid.

The main difference, though certainly not the only one, between linear and switching power supplies concerns their operating frequency: the former are necessarily tied to the mains frequency (50 or 60 Hz, depending on the country), while the latter typically operate at much higher frequencies, generally well outside the audio band.

This factor has implications primarily for the spectral distribution of noise, but the modulation effect of the mains line—which depends precisely on the power drawn by the load—remains a common consequence for both types of power supplies.

From this perspective, we can easily divide the components into two distinct groups: the first includes those with low power consumption, such as sources and preamplifiers, while the second certainly includes power amplifiers.

Using simple proportional reasoning, it becomes quite clear that power devices can have a significant influence on signal devices, while the opposite seems far less likely.

The next question seems obvious: What are the consequences of this mutual influence among the components of the audio system?

The current demand from a power amplifier results in a power draw that is partly fixed and dependent on the amplifier’s operation itself (for example, the recharging of smoothing capacitors, which occurs 100 or 120 times per second), and partly variable depending on the instantaneous power delivered to the speakers, with a strong correlation to the musical signal.

In an ideal situation where the power supply line were virtually perfect, we could probably disregard any negative impact on the final result. However, nothing is perfect, and the power supply network is subject to physical limitations that determine its behavior, moment by moment, in terms of its ability to deliver all the current required by the system.

The direct consequence of these physical limitations, combined with the absorption characteristics of the power amplifiers, results in a modulation of the mains voltage that — since it affects the entire system — is present on the secondary windings of all connected devices: sources, streamers, DACs, and preamplifiers, which will inevitably suffer negative effects.

The level of signal contamination — the only factor that ultimately matters to us — depends on various factors related to the intrinsic characteristics of the individual components and those of the power grid supplying them. However, we can conclude with a statement that is difficult to refute: no device is completely immune to its power source, and no power grid is capable of providing a perfectly stable mains voltage free from distortion, modulation, and noise.

Now let’s get back to the original question

Does it really make sense to separate the analog power supply from the digital one?

Given that most interference is generated by the power amplifiers and propagated to the other components of the system, the first step should focus precisely on these two types of components.

As for the signal electronics, the distinction between digital and analog signals appears to be less significant than the type of power supply used, which is often a determining factor in reducing background noise, a factor of fundamental importance in preserving the integrity of the musical signal, whether in digital or analog format.